Medium- and long-term outcomes of early childhood education: experiences from Turkish large-scale assessments

The paper investigates whether participation in early childhood education (ECE) is associated with improved academic outcomes in Türkiye in both the medium term (end of lower secondary/8th grade) and the long term (end of upper secondary/12th grade). The context is that ECE supports cognitive, social, and emotional development and is viewed as a lever for equity, yet Türkiye historically had low and unequal access to ECE, alongside large achievement gaps between schools and regions. The study addresses the importance of ECE for mitigating socioeconomic disparities and improving later academic performance. Research questions: (1) Is ECE participation significantly related to medium-term academic achievement at the transition to upper-secondary education? (2) Is ECE participation significantly related to long-term academic achievement and higher education transitions?

Extensive prior research generally finds positive associations between ECE and later outcomes, including higher academic achievement, reduced grade repetition, and improved cognitive and social skills (Barnett, 1995; Gorey, 2001; Campbell et al., 2002; Reynolds et al., 2003; Schweinhart et al., 2005). Large-scale programs such as Head Start show gains, often strongest for disadvantaged children, though effects may attenuate over time and some studies report nonsignificant or mixed results (Currie & Thomas, 1995; Milligan, 2012; Nold et al., 2021; Tanner et al., 2015). Internationally, ECE can reduce early achievement gaps (e.g., Indonesia’s ECED project; Jung & Hasan, 2016). In Türkiye, ECE access historically lagged OECD peers, with participation strongly linked to SES. Turkish studies—often regional or project-based—suggest ECE associates with better literacy and test outcomes and reduced inequities (Kagitcibasi et al., 2009; Ağırdağ et al., 2015; Anasız et al., 2018; Suna & Ozer, 2022), but comprehensive, representative national analyses of medium- and long-term outcomes have been limited. This study addresses that gap using nationwide administrative and assessment data.

Design: Causal-comparative (ex post facto) study comparing students who attended ECE versus those who did not to examine associations with later achievement. Population: All students who completed upper-secondary education between 2018–2021 and took the national higher education exam (YKS). The Ministry of National Education (MoNE) provided official records for 5,625,798 students. Data sources: Official MoNE e-school records including demographics (province, gender, parental education, family income) and academic indicators. ECE attendance includes both public and private institutions implementing a common national ECE curriculum. Data are not publicly available and were provided under MoNE permission (Official Letter No: E-65968543-622.03-62841376). Timing: ECE participation dated 2003–2008; medium-term achievement (LGS for upper-secondary transition) 2014–2016; long-term achievement (TYT and AYT for higher education transition) 2018–2020. Thus, the same student population has ECE, medium-term, and long-term data points. Assessments: LGS (end of grade 8) as medium-term; TYT (basic proficiency/aptitude) and AYT (field proficiency: numerical, verbal, equal weight) as long-term. Psychometric properties indicate high reliability and acceptable difficulty/discrimination across years (e.g., LGS KR-20≈0.96–0.97; TYT/AYT subtest reliabilities ~0.77–0.96). Analytical approach: Group mean comparisons between ECE attendees and non-attendees using one-way ANOVA; effect sizes via partial eta-squared (η²). Given very large N, effect size interpretation complements statistical significance. Additional long-term indicators analyzed descriptively: overall placement into higher education, placement into prestigious programs (medicine, law, dentistry), program type (open, associate, formal bachelor), and placement into top preferences. SES index constructed from parental education (mother and father separately; 1–5 each) and family income (1–5), yielding SES scores 3–15; categorized via cumulative frequency into low, medium, and high SES. SES-based subgroup comparisons explored how ECE-attendance associations vary by SES.

- Medium-term outcomes (LGS, end of grade 8): ECE attendees scored about 35 points higher than non-attendees, indicating a significant but small association (partial η² ≈ 0.03). This suggests a stronger medium-term relationship. - Long-term outcomes (TYT/AYT, higher education transition): ECE attendees had higher TYT mean scores with a small effect (η² ≈ 0.01). On AYT, differences were smaller (η² ≤ 0.01), and the AYT verbal difference was negligible. Overall, long-term associations persist but attenuate relative to medium-term. - Higher education placement rates: Overall placement into any higher education program was similar (Did not attend: 36.38%; Attended: 36.95%). - Prestigious programs: ECE attendees had higher placement into prestigious programs (medicine, law, dentistry): 13.54% (attended) vs 10.74% (did not attend). - Program types: ECE attendees were less likely to enter open and associate degree programs and more likely to enter formal bachelor programs (Formal bachelor: 61.39% attended vs 52.12% did not attend; Open programs: 3.18% attended vs 5.65% did not; Associate: 35.43% attended vs 42.23% did not). - Preferences: ECE attendees were more often placed into their top choices (e.g., First preference: 23.43% attended vs 22.45% did not; Top five preferences: 60.73% vs 57.22%), with differences roughly 1–4 percentage points across the first through first-five preferences. - SES heterogeneity: Medium-term benefits were largest for low- and middle-SES students, bringing their scores closer to the overall mean (e.g., low-SES non-attendees around −0.41 SD from mean vs ECE attendees around −0.17 SD). At the higher-education stage, differences by SES were smaller overall, indicating attenuation. Overall, ECE shows particularly beneficial associations for socioeconomically disadvantaged students, especially in the medium term.

The study’s findings indicate that ECE participation is positively associated with academic performance at both the medium and long term, answering the research questions affirmatively. The association is stronger by the end of lower secondary (LGS), consistent with developmental theory and prior evidence that ECE builds foundational skills that translate into early schooling gains, with some attenuation over time. Despite similar overall transition rates to higher education, ECE attendance is linked to more favorable higher education outcomes—greater likelihood of entering formal bachelor programs, prestigious fields, and placement into preferred programs—suggesting qualitative advantages in postsecondary trajectories. SES analyses underscore equity implications: ECE particularly benefits low- and middle-SES students in medium-term outcomes, supporting the view that universalizing quality ECE can help narrow achievement gaps in Türkiye, a system with historically large school-level disparities and late ECE expansion. The results align with international evidence on ECE’s sustained yet diminishing long-term associations and its stronger effects among disadvantaged groups.

This nationwide analysis of more than 5.6 million Turkish students links ECE attendance to higher medium-term achievement and to modest but meaningful long-term advantages, especially in entry to prestigious and formal bachelor programs and alignment with students’ preferences. Benefits are most pronounced for disadvantaged students in the medium term, supporting policies that expand equitable access to quality ECE. The study contributes representative evidence from Türkiye, where ECE has only recently scaled. Future research should incorporate more detailed controls (e.g., SES and demographic variables), examine ECE program type and duration, and employ controlled or experimental designs to strengthen causal inference and unpack mechanisms. Continued monitoring of cohorts, akin to major international ECE studies, can guide ongoing policy efforts.

- The population includes students who took Türkiye’s higher education exam between 2018 and 2020 with complete demographic data; students outside this group or not fully captured in the e-school system are excluded. - ECE program type (school- vs home-based) and duration (1–3 years) were not analyzed, limiting granularity. - SES was measured via parental education and a subjective, teacher-approved family income category; broader SES indicators were unavailable. - Administrative data are not publicly shareable; analyses rely on MoNE-provided official records. - As an ex post facto design, findings are associative; unmeasured confounders may remain.